Insurance coverage for prescription drugs is an important concern of the Medicare population, particularly for the elderly with chronic illness. Proponents of an outpatient Medicare drug benefit argue that drug coverage can offset related non-drug medical spending by increasing access to essential and newer medicines thereby reducing the need for hospitalization and institutionalization. Surprisingly, little empirical evidence supports this conventional wisdom. The purpose of this dissertation is to investigate the impact of prescription drug coverage and use of innovative pharmaceuticals on non-drug medical spending and functional disability in the elderly Medicare population. The study develops a theoretical model of how chronic illness affects the demand for medical care, which explains the tradeoff between the cost of drug utilization and the benefits of reduced risk of adverse medical events and lower functional disability. The model predicts that drug coverage and innovative drugs should reduce the probability of any acute medical spending and slow the progression of functional disability. These hypotheses will be tested using data from the 1994-1999 Medicare Current Beneficiary Survey and the National Drug Data File. The empirical analysis estimates the effect of different drug coverage types (employer-sponsored, individually-purchased Medigap, Medicare HMO, Medicaid, VA, and other public) on three outcomes: 1) drug utilization, particularly use of innovative drugs; 2) types of non-drug medical spending, including ambulatory, acute, and post-acute care; and 3) transitions in functional disability. One and two-part models of medical utilization are specified and estimated on the pooled sample, controlling for drug-related problems, individual, and area characteristics. Potential selection bias from the endogeneity of drug coverage will be addressed using several alternative strategies. The results have important policy implications for the Medicare program and health policy in general. [unreadable] [unreadable]